Skew mitigation - CS 591 K1: Data Stream Processing and Analytics Spring 2020ln 2 + O(1), with high probability • when d>2, the maximum load keeps decreasing, but only by a constant factor 10 • Consider the problem of throwing n balls to n bins sequentially (balls -> records0 码力 | 31 页 | 1.47 MB | 1 年前3- Cardinality and frequency estimation - CS 591 K1: Data Stream Processing and Analytics Spring 2020max(COUNT[j], r) R = average(COUNT) // average of all j counters output a * m * 2R // a is a constant, a 0.39701, for m 64. ≈ ≥ 12 ??? Vasiliki Kalavri | Boston University 2020 Why LogLog? Let’s0 码力 | 69 页 | 630.01 KB | 1 年前3
共 2 条
- 1